Geneious SNP detection parameter setting

[arkilis]

Hey Arkilis,

Geneious is a commercial tool. So, i dont think much people use it.

But, the option parameters are more or less same in all available variant calling tools.

Say, for example VarScan. just have a look into it.

That is a good idea.

[vishnuamaram]

Hey Arkilis,

Geneious is a commercial tool. So, i dont think much people use it.

But, the option parameters are more or less same in all available variant calling tools.

Say, for example VarScan. just have a look into it.

[arkilis]

I'm quite in it those days, but I'm not using Geneious. However, this is what I can see from the printscreen above:

1. Minimum coverage: I would set it higher, perhaps around 10. That's the number of times your SNPs will be covered. Logically, if you have only two reads covering the SNPs, that chance that both reads are errors is quite bigger than if you had 20 reads covering the same base.

2. Minimum Variant Frequency: for every SNPs that your program will find, you will have a certain number of reads that match the reference sequence, and a certain number of reads that match the alternative (SNP) call. Minimum Variant Frequency of 0.25 means that you have 1 alternative call and 3 reference calls (1/4=0.25). You can set it low in the beginning, but keep in mind that higher values are "better" for accurate calls.

3. Maximum Variavnt p-value: you control the risk of the call being made purely by chance. Go logically and set it according to the risk you want to take.

I think the rest of the options speak for itself.

Thanks for your reply.

1. minimum coverage. I think it is more depended on platform. For NGS data, you are right. 10 reads at least is more secured. For Sanger, I think two might be enough. Since Sanger data always has a good quality.

2. What you saying makes much sense.

3. HARD PART! since I don't know a proper value here. If I decrease the value, less 'SNPs' will come out. Hope those who play Geneious can see this post and provide with a moderate value..

Thanks!

[ThePresident]

I'm quite in it those days, but I'm not using Geneious. However, this is what I can see from the printscreen above:

1. Minimum coverage: I would set it higher, perhaps around 10. That's the number of times your SNPs will be covered. Logically, if you have only two reads covering the SNPs, that chance that both reads are errors is quite bigger than if you had 20 reads covering the same base.

2. Minimum Variant Frequency: for every SNPs that your program will find, you will have a certain number of reads that match the reference sequence, and a certain number of reads that match the alternative (SNP) call. Minimum Variant Frequency of 0.25 means that you have 1 alternative call and 3 reference calls (1/4=0.25). You can set it low in the beginning, but keep in mind that higher values are "better" for accurate calls.

3. Maximum Variavnt p-value: you control the risk of the call being made purely by chance. Go logically and set it according to the risk you want to take.

I think the rest of the options speak for itself.

[arkilis]

Is there anyone here using Geneious. Not quite sure about the parameters setting on "Find Variations/SNPs." Here is what I did in my project.

Any suggestion is welcome. thx